As network structures of power systems become more and more complicated, there is huge difficulty in calculation of analysis and control of a whole power system network. It is an effective method to decrease the difficulty in calculation with which the power system is partitioned into a plurality of partitions each of which is simple in structure according to analysis of the structure network of the power system. In the conventional partition method for the power systems, on the one hand, when the power system is modeled, quasi-steady characteristics are not took into account, thus leading to inaccuracies in modeling; on the other hand, the number of partitions is determined by users due to a lack of research on methods for determining the number of partitions, resulting in inaccuracy and difficulty in practical application.